fix(#614): game total score exceeding max (level1 + levelTemp double-count) by atlas-apex · Pull Request #615 · me2resh/apexyard

atlas-apex · 2026-06-09T14:15:12Z

Summary

Fixes the game total score exceeding the max (you caught it: the share text read "1188 / 1100", and the breakdown summed to 922, not the shared total).
Root cause — two levels called addScore() per round (so state.score got the sum of the rounds) while recording only the average in state.levelScores (what the results breakdown shows):
- level1 (Token splitter) — 3 rounds × per-round addScore(rs), recorded avg. Over-added ~200.
- levelTemp (Temperature dial) — per-round addScore(pts), recorded avg. Over-added ~67. Plus a double-division bug: pts = rs/3 and then avg = totalScore/SCEN.length, which capped the level at ~33/100 even on a perfect play (why it showed 33).
- 200 + 67 ≈ the 266 gap between 1188 (state.score) and 922 (Σ levelScores).
Fix — drop the per-round addScore; call addScore(avg) once at each level's finish (after levelScores is set), so every level contributes exactly its recorded 0–100 (the pattern level6 already uses). In levelTemp, accumulate totalScore += rs (not rs/3) so the average is true and a perfect run scores 100. Net: total maxes at 1100 and matches the breakdown.

Testing

node --check on the inline script — clean; LEVELS still 11; addScore is now called once per level (11 calls + the definition).
No per-round addScore(rs|pts|roundScore) remain; the pts double-division variable is removed.
Play-through: total can no longer exceed 1100; Temperature scores up to 100 on a correct/in-band answer.

Closes #614

Glossary

Term	Definition
`state.score`	The running total shown in the score pill + share text.
`state.levelScores[]`	Per-level scores in the results breakdown (last value per level).
Over-count	`state.score` receiving more additions than the per-level scores sum to.

state.score could exceed the 1100 max (observed share text "1188/1100") and didn't match the per-level breakdown. Two levels called addScore() per round (summing into state.score) but recorded only the average in levelScores: - level1 (Token splitter): drop the per-round addScore(rs); call addScore(avg) once at finish, after levelScores is set — so the level contributes its recorded 0-100 exactly once (matching every other level / level6's pattern). - levelTemp (Temperature dial): same per-round fix, plus fix a double-division (pts = rs/3 then avg = totalScore/SCEN.length capped the level at ~33 even on a perfect play). Now accumulate totalScore += rs so avg is the true average and a perfect run scores 100. Result: every level adds exactly its 0-100 to state.score once; total maxes at 1100 and matches the breakdown. node --check clean; LEVELS still 11. Closes #614 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

atlas-apex

Code Review: PR #615

Commit: 3bf04a434a1a4d08441fa6b5f30872278cf96682

Summary

Fixes a scoring bug in site/game.html where the game total could exceed the 1100 max (share text showed "1188/1100") and didn't match the per-level breakdown. Two levels called addScore() per round (summing rounds into state.score) while recording only the average in state.levelScores. The fix drops the per-round adds and calls addScore(avg) once at each level's finish — the pattern the other 9 levels already use. levelTemp also had a double-division (pts = rs/3 then avg = totalScore/SCEN.length) that capped the level at ~33 on a perfect play; now totalScore += rs so the average is true. Diff: +3/-5, single file.

Verification

node --check on the inline script: clean. One inline <script> block; parses without error.
LEVELS still 11 — render entries: level1, level2, levelTemp, level3, levelCutoff, level4, level5, level6, level7, levelCost, levelLoop.
The invariant holds across all 11 levels. Every state.levelScores[state.level] = X write is immediately followed by addScore(X) with the identical variable — verified by pairing all 11 sites:
- level1@1273 avg → addScore@1274 avg
- level2@1360 sc → @1361 sc
- levelTemp@2245 avg → @2246 avg
- level3@1522 sc → @1523 sc
- levelCutoff@1704 sc → @1705 sc
- level4@1803 sc → @1804 sc
- level5@2090 sc → @2091 sc
- level6@1930 totalScore → @1931 totalScore (legitimately once = its levelScores; fine)
- level7@2343 sc → @2344 sc
- levelCost@2475 sc → @2476 sc
- levelLoop@2600 sc → @2601 sc
- No level adds per-round or adds a value different from what it records. addScore does a bare state.score += n (no clamp), so state.score === Σ levelScores, max = 11 × 100 = 1100. Defect fixed.
levelTemp: pts fully removed (the only remaining pts is the literal UI string 'SKIP? (0 pts)', not the variable — no dangling reference). totalScore += rs correct; rs declared in the lock handler and in scope where used. Perfect in-band play → every round rs=100 → avg = Math.round(totalScore/SCEN.length) = 100 (no longer 33). addScore(avg) is in the finish branch.
level1: rs still computed (Math.round((correct/real.size)*100 - wrong*15 - missed*12), clamped 0–100) and totalScore += rs per round retained — the removed line was only addScore(rs). rs still used for the per-round comparison display. avg = Math.round(totalScore / ROUNDS.length) unchanged. addScore(avg) is in the finish else branch, not per round.
No other behavior changed; diff matches the stated +3/-5.

Checklist Results

✅ Architecture & Design: Pass (static single-file change, no layering concern)
✅ Code Quality: Pass (minimal, surgical, removes the double-division dead path)
✅ Testing: Pass (node --check is the project's gate for this static site; clean. No test suite by design)
✅ Security: Pass (no auth/secrets/user-data; pure client-side scoring)
✅ Performance: Pass (no hot-path or query changes)
✅ PR Description & Glossary: Pass (Summary + Testing + Glossary all present; narrative bullets)
✅ Summary Bullet Narrative: Pass (each bullet states what + why)
✅ Technical Decisions (AgDR): N/A (bug fix, no library/architecture/pattern decision)
✅ Adopter Handbooks: N/A (no architecture/general/language violations triggered)

Issues Found

None.

Handbook Findings

Migration Safety (blocking): not triggered — no migration files in the diff.
Clean Architecture Layers (advisory): N/A — single static HTML file, no cross-layer imports.
Commit Message Quality (advisory): Pass — commit 3bf04a4 has a descriptive WHY-bearing subject and a body explaining problem, root cause, the two fixes with rationale, result, and verification. Single Closes #614, co-authorship line present.

Suggestions

None blocking. (Optional, for a future PR: the per-level scoring pattern is now uniform — a one-line console.assert(state.score === state.levelScores.reduce(...)) dev-only guard could lock the invariant in if the game grows past 11 levels. Not needed for this fix.)

Verdict

APPROVED

PR body has Summary + Testing + Glossary; Closes #614 (issue OPEN); commit SHA matches HEAD (3bf04a4...). Base is dev, single file, +3/-5.

🤖 Reviewed by Rex (Code Reviewer Agent)
📌 Reviewed commit: 3bf04a434a1a4d08441fa6b5f30872278cf96682

atlas-apex

Code Review: PR #615

Commit: 3bf04a434a1a4d08441fa6b5f30872278cf96682

Summary

Fixes a scoring bug in site/game.html where the game total could exceed the 1100 max (share text showed "1188/1100") and didn't match the per-level breakdown. Two levels called addScore() per round (summing rounds into state.score) while recording only the average in state.levelScores. The fix drops the per-round adds and calls addScore(avg) once at each level's finish — the pattern the other 9 levels already use. levelTemp also had a double-division (pts = rs/3 then avg = totalScore/SCEN.length) that capped the level at ~33 on a perfect play; now totalScore += rs so the average is true. Diff: +3/-5, single file.

Verification

node --check on the inline script: clean. One inline <script> block; parses without error.
LEVELS still 11 — render entries: level1, level2, levelTemp, level3, levelCutoff, level4, level5, level6, level7, levelCost, levelLoop.
The invariant holds across all 11 levels. Every state.levelScores[state.level] = X write is immediately followed by addScore(X) with the identical variable — verified by pairing all 11 sites:
- level1@1273 avg → addScore@1274 avg
- level2@1360 sc → @1361 sc
- levelTemp@2245 avg → @2246 avg
- level3@1522 sc → @1523 sc
- levelCutoff@1704 sc → @1705 sc
- level4@1803 sc → @1804 sc
- level5@2090 sc → @2091 sc
- level6@1930 totalScore → @1931 totalScore (legitimately once = its levelScores; fine)
- level7@2343 sc → @2344 sc
- levelCost@2475 sc → @2476 sc
- levelLoop@2600 sc → @2601 sc
- No level adds per-round or adds a value different from what it records. addScore does a bare state.score += n (no clamp), so state.score === Σ levelScores, max = 11 × 100 = 1100. Defect fixed.
levelTemp: pts fully removed (the only remaining pts is the literal UI string 'SKIP? (0 pts)', not the variable — no dangling reference). totalScore += rs correct; rs declared in the lock handler and in scope where used. Perfect in-band play → every round rs=100 → avg = Math.round(totalScore/SCEN.length) = 100 (no longer 33). addScore(avg) is in the finish branch.
level1: rs still computed (Math.round((correct/real.size)*100 - wrong*15 - missed*12), clamped 0–100) and totalScore += rs per round retained — the removed line was only addScore(rs). rs still used for the per-round comparison display. avg = Math.round(totalScore / ROUNDS.length) unchanged. addScore(avg) is in the finish else branch, not per round.
No other behavior changed; diff matches the stated +3/-5.

Checklist Results

✅ Architecture & Design: Pass (static single-file change, no layering concern)
✅ Code Quality: Pass (minimal, surgical, removes the double-division dead path)
✅ Testing: Pass (node --check is the project's gate for this static site; clean. No test suite by design)
✅ Security: Pass (no auth/secrets/user-data; pure client-side scoring)
✅ Performance: Pass (no hot-path or query changes)
✅ PR Description & Glossary: Pass (Summary + Testing + Glossary all present; narrative bullets)
✅ Summary Bullet Narrative: Pass (each bullet states what + why)
✅ Technical Decisions (AgDR): N/A (bug fix, no library/architecture/pattern decision)
✅ Adopter Handbooks: N/A (no architecture/general/language violations triggered)

Issues Found

None.

Handbook Findings

Migration Safety (blocking): not triggered — no migration files in the diff.
Clean Architecture Layers (advisory): N/A — single static HTML file, no cross-layer imports.
Commit Message Quality (advisory): Pass — commit 3bf04a4 has a descriptive WHY-bearing subject and a body explaining problem, root cause, the two fixes with rationale, result, and verification. Single Closes #614, co-authorship line present.

Suggestions

None blocking. (Optional, for a future PR: the per-level scoring pattern is now uniform — a one-line console.assert(state.score === state.levelScores.reduce(...)) dev-only guard could lock the invariant in if the game grows past 11 levels. Not needed for this fix.)

Verdict

APPROVED

PR body has Summary + Testing + Glossary; Closes #614 (issue OPEN); commit SHA matches HEAD (3bf04a4...). Base is dev, single file, +3/-5.

🤖 Reviewed by Rex (Code Reviewer Agent)
📌 Reviewed commit: 3bf04a434a1a4d08441fa6b5f30872278cf96682

github-actions Bot added github-issue has-ticket labels Jun 9, 2026

atlas-apex commented Jun 9, 2026

View reviewed changes

atlas-apex merged commit 2224a53 into dev Jun 9, 2026
7 checks passed

atlas-apex deleted the fix/GH-614-game-score branch June 9, 2026 14:19

This was referenced Jun 9, 2026

[Chore] Release v3.1.3 — game scoring fix #616

Closed

release(#616): v3.1.3 #617

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(#614): game total score exceeding max (level1 + levelTemp double-count)#615

fix(#614): game total score exceeding max (level1 + levelTemp double-count)#615
atlas-apex merged 1 commit into
devfrom
fix/GH-614-game-score

atlas-apex commented Jun 9, 2026

Uh oh!

atlas-apex left a comment

Uh oh!

atlas-apex left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

atlas-apex commented Jun 9, 2026

Summary

Testing

Glossary

Uh oh!

atlas-apex left a comment

Choose a reason for hiding this comment

Code Review: PR #615

Summary

Verification

Checklist Results

Issues Found

Handbook Findings

Suggestions

Verdict

Uh oh!

atlas-apex left a comment

Choose a reason for hiding this comment

Code Review: PR #615

Summary

Verification

Checklist Results

Issues Found

Handbook Findings

Suggestions

Verdict

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants